llmrun

NVIDIA · Turing

Best AI Models for NVIDIA GeForce GTX 1650 SUPER (4 GB)

Runs only small models well; 7–9B doesn't fit in 4 GB.

4 GB GDDR6192 GB/s9 TFLOPS FP16100 WTuring 2019
Buyabout $159

4 GB at 192 GB/s: small models only (~3B at Q4). No Tensor Cores.

What LLMs can the GTX 1650 SUPER run?

#ModelSpeedllmrun ScoreQuant · VRAM
1
Qwen3.5 4BAlibabaHow to run →
Q4_K_M · 3.4 GB · 38.3 t/s
38.3 tok/sQ4_K_M · 3.4 GB
2
MiniCPM5 2BOpenBMB
Q4_K_M · 2.5 GB · 65.7 t/s
65.7 tok/sQ4_K_M · 2.5 GB
3
Gemma 4 E2B ITGoogle
Q4_K_M · 3.5 GB · 36.4 t/s
36.4 tok/sQ4_K_M · 3.5 GB
4
LFM3 1.7BLiquid AI
Q4_K_M · 2.1 GB · 86.7 t/s
86.7 tok/sQ4_K_M · 2.1 GB
5
LFM3 Omni 2BLiquid AI
Q4_K_M · 2.4 GB · 73.4 t/s
73.4 tok/sQ4_K_M · 2.4 GB
See all 151 models on the GTX 1650 SUPER →Best quant that fits · speeds estimated on this card

About the GTX 1650 SUPER

The GTX 1650 SUPER is a solid mid-range card in the 4–6 GB class: it runs only small models well. It is the most efficient per watt in its tier, and the best speed per dollar of the cards compared at ≈$99 used. Turing generation, released 2019.

Speed in 4–6 GB#3192 GB/s
Per watt#10.35 tok/s per W on Qwen3.5 9B (9B params)
Per dollar#135.2 tok/s per $100 on Qwen3.5 9B (9B params)
GenerationTuringreleased 2019

GTX 1650 SUPER strengths and weaknesses

Strengths

  • Low power draw: 100 W board power; a 200 W PSU is enough

Weaknesses

  • 7–9B models don't fit: Qwen3.5 9B needs 6.6 GB at Q4 vs 4 GB here
  • Slow prompt reading: A 4K-token prompt on Qwen3.5 4B waits 9.6 s for the first token
  • No FP8 or NVFP4: Turing predates both; use GGUF, AWQ or GPTQ builds

Speed per dollar and per watt

On Qwen3.5 9B (9B params) at Q4_K_M. New prices are compared with new, used with used. The last two columns are the speed you get per $100 of price and per 100 W of board power; higher is better. The GTX 1650 SUPER has no new price, so the New view shows it at its MSRP.

CardNew pricetok/stok/s per $100tok/s per 100 W
GTX 1650 SUPER4 GB
about $15934.8
21.9
34.8
RTX 50608 GBBest new value
about $46045.8
10.0
31.6
RTX 507012 GB
about $84068.7
8.2
27.5
RX 907016 GB
about $80060.4
7.6
27.5
RTX 40608 GB
about $52027.8
5.3
24.2
CardUsed pricetok/stok/s per $100tok/s per 100 W
GTX 1650 SUPER4 GB
about $9934.8
35.2
34.8
RTX 3060 12GB12 GB
about $32936.8
11.2
21.6
RTX 50608 GBBest used value
about $43045.8
10.7
31.6
RX 907016 GB
about $65060.4
9.3
27.5
RTX 407012 GB
about $57551.5
9.0
25.8
RTX 40608 GB
about $33527.8
8.3
24.2
RTX 4070 SUPER12 GB
about $67551.5
7.6
23.4
RTX 4060 Ti 8GB8 GB
about $40029.4
7.4
18.4

Cost per token: the GTX 1650 SUPER against 4–16 GB cards

On Qwen3.5 9B (9B params, Q4_K_M, 4K context). Speed, cost per token and power per token depend on the model.

4 GB12 GB16 GBthis cardat MSRP
$0.00$0.50$1.00$1.50$0$1,000$ per 1M output tokens · lower is betterused price →RX 9070 · 16 GB · ≈$650 · $0.83 per 1M tokensRTX 3060 12GB · 12 GB · ≈$329 · $0.76 per 1M tokensRTX 4060 Ti 8GB · 8 GB · ≈$400 · $1.09 per 1M tokensRTX 5060 · 8 GB · ≈$430 · $0.73 per 1M tokensRTX 5070 · 12 GB · ≈$736 · $0.83 per 1M tokensRTX 4070 SUPER · 12 GB · ≈$675 · $1.01 per 1M tokensRTX 4070 Ti · 12 GB · ≈$722 · $1.12 per 1M tokensRTX 4060 · 8 GB · ≈$335 · $0.94 per 1M tokensRTX 4070 · 12 GB · ≈$575 · $0.87 per 1M tokensGTX 1650 SUPER · 4 GB · ≈$99 · $0.30 per 1M tokensGTX 1650 SUPERRTX 5060RTX 5070RX 9070RTX 4070 SUPERRTX 4070 TiRTX 4070
$0.00$0.50$1.00$1.50$0$1k$ per 1M output tokens · lower is betterused price →RX 9070 · 16 GB · ≈$650 · $0.83 per 1M tokensRTX 3060 12GB · 12 GB · ≈$329 · $0.76 per 1M tokensRTX 4060 Ti 8GB · 8 GB · ≈$400 · $1.09 per 1M tokensRTX 5060 · 8 GB · ≈$430 · $0.73 per 1M tokensRTX 5070 · 12 GB · ≈$736 · $0.83 per 1M tokensRTX 4070 SUPER · 12 GB · ≈$675 · $1.01 per 1M tokensRTX 4070 Ti · 12 GB · ≈$722 · $1.12 per 1M tokensRTX 4060 · 8 GB · ≈$335 · $0.94 per 1M tokensRTX 4070 · 12 GB · ≈$575 · $0.87 per 1M tokensGTX 1650 SUPER · 4 GB · ≈$99 · $0.30 per 1M tokensGTX 1650 SUPERRTX 5060RTX 5070
Cost per 1M output tokenson Qwen3.5 9B · card over 3 years + power
$0.30
Card, used price
$0.18
Power
$0.12
3-year cost of ownership$165
Or rent an RTX 3080 Ti in the cloud$0.13/hr

Renting costs as much as buying after ≈800 hours (≈0.5 years at 4 h a day).

Include card price

Each dot is a card; hover or tap it for its numbers, click to open it. Cards without a price for the chosen basis are left out. How we calculate cost per token

Similar GPUs

Cards people weigh against the GTX 1650 SUPER, each with the one reason it is worth a look. Speeds on Qwen3.5 9B.

8 GB more and 6% fasterRTX 3060 12GBNVIDIA · 12 GB · Ampere
▲ 6% speed▲ 8 GBabout $329rent $0.06/hr
12 GB more and 74% fasterRX 9070AMD · 16 GB · RDNA 4
▲ 74% speed▲ 12 GBabout $800
4 GB more, but slowerRTX 4060 Ti 8GBNVIDIA · 8 GB · Ada Lovelace
▼ 15% speed▲ 4 GBabout $399rent $0.15/hr
2 GB more and 14% fasterGTX 1660 SUPERNVIDIA · 6 GB · Turing
▲ 14% speed▲ 2 GBabout $229
4 GB more and 16% fasterA750Intel · 8 GB · Alchemist
▲ 16% speed▲ 4 GBabout $289
4 GB more and 16% fasterA580Intel · 8 GB · Alchemist
▲ 16% speed▲ 4 GBabout $179

GTX 1650 SUPER tokens per second by model size

Output speed depends on memory bandwidth (192 GB/s). Grey rows need more than 4 GB.

  1. 9BQwen3.5 9B · not at 4-bit
    needs 6.6 GB
  2. 14BQwen3 14B · not at 4-bit
    needs 11.9 GB
  3. 27–31BQwen3.8 27B · not at 4-bit
    needs 17.9 GB
  4. 35B MoEQwen3.6 35B A3B · not at 4-bit
    needs 22.0 GB
  5. 70BLlama 3.3 70B Instruct · not at 4-bit
    needs 51.3 GB
  6. 120B MoEGPT OSS 120B · not at 4-bit
    needs 70.8 GB

Estimated speeds: Q4_K_M, 4K context, one user. How we estimate

All 562 models on the GTX 1650 SUPER

Featured only

Best quant that fits, ordered by llmrun Score. Fit grades how well it runs here, S best. How ranking works

LLM models compatible with NVIDIA GeForce GTX 1650 SUPER — ranked by Score, recency and Fit
ModelVRAMScoreFit
Qwen3.8 4B Distill
Q4_K_M·38.3 t/s
3.4 GBA67, Runs well
Llama 3.2 1B Instruct
Q4_K_M·152.2 t/s
2.3 GBS97, Runs great
Tiny Random OPTForCausalLM
BF16·12480.0 t/s
0.0 GBS100, Runs great
LFM2.5 1.2B JP 202606
Q4_K_M·116.6 t/s
1.2 GBS94, Runs great
Wald 4B
Q4_K_M·38.3 t/s
3.4 GBA67, Runs well
The GuageLLM 23M
Q4_K_M·6240.0 t/s
0.0 GBS100, Runs great
SmolVLM 500M Instruct
Q4_K_M·180.9 t/s
0.9 GBS99, Runs great
Moondream2
Q4_K_M·98.3 t/s
3.0 GBS91, Runs great
Moore Lm
BF16·1386.7 t/s
0.7 GBS100, Runs great
LLaMA3.1 8B Instruct DFlash UltraChat
BF16·51.1 t/s
2.7 GBA80, Runs well
SmolVLM 256M Instruct
Q4_K_M·249.6 t/s
0.6 GBS100, Runs great
Lumi 4B
Q4_K_M·43.0 t/s
2.9 GBA77, Runs well
MiniCPM V 4
Q4_K_M·44.9 t/s
3.1 GBA77, Runs well
Opt 125M
BF16·445.7 t/s
0.3 GBS100, Runs great
LFM2.5 VL 450M
Q4_K_M·195.0 t/s
0.8 GBS100, Runs great

GTX 1650 SUPER specs, power and PSU

Specs

Architecture
Turing
VRAM
4 GB GDDR6
Memory bandwidth
192.0 GB/s
FP16 (dense)
8.83 TFLOPS
CUDA cores
1,280
Tensor cores
0
Compute capability
7.5
TDP
100 W
Released
2019-11-22
MSRP
$159

GTX 1650 SUPER FAQ

Is the GTX 1650 SUPER still worth it in 2026?

At ≈$99 used you get 35.2 tok/s per $100 on Qwen3.5 9B, which ranks 1st of the 8 cards compared here; it is also the 3rd fastest out of 8 cards with 4–6 GB. It runs only small models well; 7–9B doesn't fit in 4 GB.

What is the largest model the GTX 1650 SUPER can run?

Qwen3.5 4B (4.7B) is the largest at 4-bit or better: at Q4_K_M it uses 3.4 GB of 4 GB, and runs at about 38.3 tok/s.

Can the GTX 1650 SUPER run 7–9B models?

No. Qwen3.5 9B needs 6.6 GB at Q4, more than the 4 GB on this card.

How many tokens per second does the GTX 1650 SUPER give?

About 38.3 tok/s on Qwen3.5 4B at Q4_K_M, with the first token after 9.6 s on a 4K-token prompt.

What power supply does the GTX 1650 SUPER need?

Plan a 200 W PSU (2× the 100 W TDP).

Is renting cheaper than buying a GTX 1650 SUPER?

An RTX 3080 Ti (12 GB) rents from $0.13/hr. Renting costs as much as buying after ≈800 hours (≈0.5 years at 4 h a day).

Related

Best coding LLMs8 GB VRAMGPUs with 4 GB or moreRTX 3070 TiA770 8GBRTX 2080 SUPER